vectordotdev/vector · error · io::Error

InvalidInput

InvalidInput

Error message

cannot extend a file through the truncation API

What it means

In the disk-buffer v2 writer, `truncate_current_data_file` applies a shrink to the current data file after flushing, and guards the invariant that the requested size must not exceed `writer.current_data_file_size`. Violating it returns `InvalidInput` "cannot extend a file through the truncation API". Because this runs after a writer-checkpoint scan (with the reader dropped for Windows compatibility), an oversized request indicates the writer's notion of the file size disagrees with the truncation target — an internal consistency failure, not normal operation.

Source

Thrown at lib/vector-buffers/src/variants/disk_v2/writer.rs:1241

                        WriterCheckpointScanAction::Continue => {}
                        WriterCheckpointScanAction::Stop => break,
                    }
                }
            }
        }

        // Windows does not permit truncating a file while the scan's read handle remains open.
        drop(reader);
        self.apply_writer_checkpoint_scan(scan).await
    }

    async fn truncate_current_data_file(&mut self, size: u64) -> io::Result<()> {
        let writer = self
            .writer
            .as_mut()
            .expect("writer should exist after `ensure_ready_for_write`");
        if size > writer.current_data_file_size {
            return Err(io::Error::new(
                io::ErrorKind::InvalidInput,
                "cannot extend a file through the truncation API",
            ));
        }
        writer.writer.flush().await?;

        let data_file_path = self.ledger.get_current_writer_data_file_path();
        self.ledger
            .filesystem()
            .truncate_file(&data_file_path, size)
            .await?;

        writer.current_data_file_size = size;
        self.data_file_size = size;

        Ok(())
    }

View on GitHub (pinned to 3708c39b12)

Solutions

1. If the buffer's on-disk state is suspect, move aside/delete the `data_dir` buffer directory and let Vector recreate it (accepted data loss of buffered events).
2. Report upstream with the Vector version and buffer config if it reproduces on an untouched data directory — the size mismatch is an internal invariant break.
3. Ensure only one Vector instance uses a given buffer directory and that versions are not interleaved across restarts.

// Internal guard before requesting truncation:
if size > writer.current_data_file_size {
    warn!(size, current = writer.current_data_file_size,
          "skipping truncation that would extend the data file");
    return Ok(());
}

When it happens

Trigger: Buffer internals requesting truncation past the tracked data-file size — e.g. corrupted ledger/checkpoint state after a crash, manual tampering with the data directory, or a bug in offset arithmetic feeding `truncate_current_data_file`. Ordinary Vector users should never see this from config alone.

Common situations: Data-disk-buffer files edited/copied by hand; mixing buffer directories between Vector versions or instances; disk state mutated while Vector was stopped; bugs in buffer checkpoint recovery.

Related errors

• InvalidInput
• effective reader file ID must be in the checkpoint window
• record was already validated
• Event count for a record cannot exceed 2^64 events.
• a record with a next ID must have an event count

AI-assisted analysis of vectordotdev/vector@3708c39b12 (2026-08-20). Data as JSON: /api/errors/f6b52ada9fe44b0b. Report an issue: GitHub.